Martin W.

Key Skills

Software

Label Studio

Surge AI

AWS SageMaker

CVAT

Data Annotation Tech

Encord

Labelbox

SuperAnnotate

V7 Labs

Appen

Argilla

Top Subject Matter

LLM evaluation

prompt engineering

safety testing

Top Data Types

Image

Text

Audio

Top Task Types

Bounding Box

Text Generation

Question Answering

Text Summarization

RLHF

Object Detection

Cuboid

Fine-tuning

Transcription

Red Teaming

Evaluation/Rating

Prompt + Response Writing (SFT)

Classification

Polygon

Segmentation

Point/Key Point

Data Collection

Polyline

Entity (NER) Classification

Function Calling

Computer Programming/Coding

Freelancer Overview

Over the past 4+ years, I have built deep expertise in AI training data operations across text, image, audio, and video modalities. I specialize in LLM evaluation, RLHF-based preference ranking, red teaming, and safety testing — having designed complex prompts, developed evaluation rubrics, and ranked model outputs against multi-dimensional quality criteria including helpfulness, harmlessness, honesty, and instruction-following accuracy. My work has directly improved model performance across 20+ client projects spanning education, technical, and conversational domains. What distinguishes me is my combination of hands-on annotation precision and quality systems thinking. I have developed annotation guidelines that improved inter-annotator agreement (Cohen's Kappa) by 25% across distributed teams of 15+ annotators, mentored and onboarded new team members (reducing ramp-up time by 30%), and built standardized quality checklists that cut content review turnaround by 30%. I am proficient across industry-standard toolchains including Label Studio, Labelbox, CVAT, and Scale AI platforms, with foundational Python and SQL skills for data validation and analysis. I hold Google's AI Essentials Certification and am currently pursuing an M.Ed. in Educational Technology — a background that uniquely positions me to understand how humans learn from and interact with AI systems.

Labeling Experience

AI Content Trainer & Prompt Evaluation Specialist (Freelance, Remote)

Don't discloseTextPrompt + Response Writing (SFT)

Delivered prompt engineering for training and fine-tuning large language models focused on reasoning, factual accuracy, and contextual coherence. Applied RLHF-style preference ranking by comparing model outputs against quality criteria such as helpfulness, harmlessness, honesty, and instruction-following. Led red teaming to uncover vulnerabilities and safety risks and documented findings with reproducible test cases. • Developed and refined annotation guidelines and evaluation rubrics to improve inter-annotator agreement. • Evaluated AI outputs for factual correctness, logical consistency, tone appropriateness, and style adherence across client projects. • Collaborated with QA specialists and ML engineers to update labeling protocols and ensure dataset quality. • Mentored and onboarded new annotators, reducing ramp-up time.

2023 - Present

AI Trainer/Data Labeler

TextRLHF

Conducted AI training and data labeling for large language models focused on educational and academic content domains. Core responsibilities included crafting high-quality, nuanced text responses to train LLMs on reasoning, clarity, and contextual understanding; evaluating AI-generated outputs for accuracy, coherence, tone consistency, and adherence to project-specific style guidelines; and performing RLHF-style preference ranking by comparing model responses and selecting optimal outputs based on predefined quality criteria. Specialized in text classification, content verification, sentiment analysis, and structured data tagging across academic writing, research methodology, and English language instruction domains. Maintained high accuracy rates throughout all annotation tasks and collaborated with quality assurance specialists to refine annotation guidelines and improve dataset consistency. Leveraged AI Essentials certification from Google's My Skills Workshop and competencies in machine learning algorithms and large language model architectures to deliver training data that directly improved model performance.

2026 - 2026

AI Trainr/Data Labeler

TextRLHF

Specialized in RLHF (Reinforcement Learning from Human Feedback) text annotation and AI model training for large language models. Craft high-quality, nuanced text responses to train and fine-tune LLMs on reasoning, clarity, and contextual understanding. Core Responsibilities: Performed RLHF-style preference ranking by comparing AI-generated responses and selecting optimal outputs based on predefined quality criteria including accuracy, coherence, and tone consistency. Evaluated AI-generated text outputs for adherence to project-specific style guidelines, ensuring training data reflects expert-level knowledge and communication standards. Crafted detailed, nuanced responses across academic and educational content domains to improve model reasoning and contextual understanding. Classified and annotated text data across multiple dimensions: content quality, grammatical correctness, semantic relevance, and educational appropriateness. Collaborated with quality assurance specialists to refine annotation guidelines and improve consistency across training datasets. Project Scope: Contributed to the training and fine-tuning of large language models through thousands of annotated text samples and preference-ranked response pairs, specializing in educational and academic content verticals. Quality Standards: Maintained high accuracy rates and strict adherence to complex annotation protocols, ensuring reliable training data for machine learning pipelines.

2025 - 2026

Content Quality Analyst — AI Training Data Preparation

OtherAudioClassification

Structured and labeled large-scale datasets for supervised learning with consistent taxonomy and metadata tagging across text, image, and audio corpora. Verified and corrected AI-generated content samples for factual accuracy, grammatical correctness, and alignment with domain knowledge bases. Managed end-to-end labeling workflows for sentiment analysis, topic categorization, content moderation, and intent classification training sets. • Developed standardized quality checklists and documentation templates to reduce review turnaround time. • Ensured consistent labeling decisions by following client-specific guidelines. • Audited content for compliance with moderation and intent taxonomy requirements. • Supported recurring engagements with improved review and verification processes.

2025 - 2025

Data Annotation & Model Evaluation Specialist (Scale AI / Remotasks)

Scale AIText

Completed 5,000+ high-precision annotations including named entity recognition and text classification. Produced bounding box and polygon-style labels and multi-label sentiment annotations while maintaining 98%+ accuracy through calibration and guideline adherence. Evaluated computer vision and NLP model outputs and delivered structured feedback for retraining priorities. • Implemented quality-focused annotation practices to reduce systematic labeling ambiguities. • Escalated edge cases to project leads to refine guidelines and lower error rates. • Participated in pilot programs for new annotation toolchains and provided usability feedback. • Contributed to consistent labeling across an annotator pool.

2022 - 2022

Education

Work History